Nvidia is investing $3.5 billion in Taiwanese chipmaker MediaTek, a deal that will give MediaTek access to Nvidia鈥檚 NVLink Fusion ecosystem for designing custom AI chips that can plug directly into Nvidia-based data centers. The move comes as cloud providers and AI labs, including Amazon, Google, Microsoft, OpenAI, and Anthropic, build their own silicon to reduce dependence on Nvidia鈥檚 GPUs. Under the arrangement, MediaTek can keep making custom chips for those customers while Nvidia maintains its position as the dominant infrastructure layer in AI data centers.
Dion Harris, Nvidia鈥檚 senior director of HPC and AI hyperscaler infrastructure solutions, told reporters that Nvidia has positioned itself as an AI infrastructure company that expanded beyond standalone computing chips years ago. He said the deal lets MediaTek鈥檚 customers standardize on Nvidia鈥檚 rack-scale architecture and deploy their custom chips alongside Nvidia hardware using the same platform. The partnership fits a pattern of Nvidia investing in companies that feed back into its own ecosystem, a circular financing approach the company has used repeatedly in recent years.
The NVLink Fusion ecosystem includes NVLink, a high-speed interconnect that lets chips from different vendors communicate quickly, even if they are not Nvidia-made. Last week, Nvidia announced a similar partnership with Amazon Web Services, though without a direct investment. AWS is set to deploy an additional 2 million Nvidia GPUs and will also integrate NVLink Fusion into its infrastructure.
MediaTek has been building its custom data center ASIC business, saying in June it expects that operation to generate $2 billion in revenue in 2026. The company has expertise in custom chips for smartphones, smart homes, automobiles, and wireless communications, which it will continue to apply alongside Nvidia on the DGX Spark desktop AI computer. The two firms are also extending their work on the RTX Spark, Nvidia鈥檚 push into consumer AI PCs, according to the announcement.
In the automotive sector, MediaTek and Nvidia will keep developing platforms for AI-powered, software-defined vehicles, a field Nvidia describes as the era of physical AI. MediaTek鈥檚 auto platforms rely on Nvidia鈥檚 RTX graphics for intelligent cockpit systems and work with Nvidia Drive AGX, Nvidia鈥檚 in-car computing platform for autonomous driving. Nvidia CEO Jensen Huang said in a statement that the companies are building platforms that bring Nvidia accelerated computing to new markets, from large-scale AI factories to PCs and cars, giving customers freedom to create differentiated AI systems at scale.
Harris noted that every major cloud provider and model builder is deploying Nvidia鈥檚 platform in some form. By partnering with MediaTek, he said, Nvidia is opening its ecosystem to MediaTek鈥檚 full customer base, allowing those companies to run custom silicon alongside Nvidia gear on the same standard infrastructure. MediaTek has not publicly detailed its customer list for custom AI chips, but the investment signals Nvidia鈥檚 intent to grow that side of MediaTek鈥檚 business as long as the chips remain compatible with Nvidia鈥檚 data center framework.
More hardware news from TechManNews.








